Student Performance Predictor

A machine learning web application to predict student academic performance based on demographic and previous scores. This project demonstrates end-to-end ML workflow: data ingestion, preprocessing, model training, evaluation, and deployment.

---

Project Overview

• Goal: Predict student scores (reading & writing) based on:
  • Gender
  • Race/Ethnicity
  • Parental education
  • Lunch type
  • Test preparation course
  • Math scores
• Approach: Multiple regression models (Random Forest, Gradient Boosting, Linear Regression, XGBoost, CatBoost, AdaBoost, Decision Tree) were trained and evaluated using R² scores to select the best model.
• Data Handling: Used pandas and numpy for preprocessing, categorical encoding, and scaling.
• Model Saving: Models serialized using dill/pickle for deployment.
• Web Interface: Flask app with a simple UI where users input student data and get predictions.

---

Key Features

• End-to-End ML Pipeline: From CSV ingestion → preprocessing → training → prediction.
• Multiple Models: Automatic selection of the best-performing model.
• Reusable Components: Separate modules for data ingestion, transformation, training, and prediction.
• Deployment Ready: Ready to deploy on free hosting platforms like Render.

---

Deployment

• Deployed URL: Student Performance Predictor
• Technology: Flask, Gunicorn
• Usage: Enter student details and click Predict to get predicted scores.

---

Screenshots sample

---

Author

Manish Raj
GitHub Profile